Maintenance policy performance assessment in presence of imprecision based on Dempster–Shafer Theory of Evidence

Abstract

To cite this version:Piero Baraldi, Michele Compare, Enrico Zio. Maintenance policy performance assessment in presence of imprecision based on Dempster-Shafer Theory of Evidence. Information Sciences, Elsevier, 2013, 245, pp.112 -131 Abstract: The aim of this work is to assess the performance of a maintenance policy when a stochastic model of the life of the component of interest is known, but relies on parameters that are imprecisely known, and only through information elicited from experts. The case in which the information used to feed the model comes from a single expert has been investigated by the authors in a previous work. This paper deals with the different situation in which a number of experts are involved in the elicitation of the uncertain parameters; in particular, each expert provides an interval he/she believes containing the unknown value of the parameter which he/she is knowledgeable about. The different type of available information calls for the development of a different method to represent and propagate the associated uncertainty. Resorting to Probability theory to address this issue is questionable. Then, a technique based on the Dempster-Shafer Theory of Evidence (DSTE) is embraced in this work, which allows facing a practical case study concerning the check valve of a turbo-pump lubricating system in a Nuclear Power Plant. The output of such method consists of couples of Lower and Upper cumulative distributions describing the uncertainty in the maintenance performance indicators of interest (i.e., unavailability and costs), which accounts for both the aleatory and epistemic contributions.

Genetic algorithms for condition-based maintenance optimization under uncertainty

“…This importance is witnessed by the large amount of literature produced on this topic (e.g., [3], [5], [6], [9], [21], [45], [46]). In other words, the maintenance models have to take into account the uncertainties affecting their parameters, which usually come from limited evidence available in the field.…”

Uncertainty treatment in expert information systems for maintenance policy assessment

“…For example, in [21] a different technique has been proposed to cope with the maintenance assessment issue in the case in which a team of experts is available to provide the ill-defined parameters, whereas the method proposed in this work assumes that there is just one expert to give them.…”

“…Furthermore, this solution leaves open the question: shall we consider the uncertainty in the membership defining the type-2 fuzzy set? A different approach to the problem of assessing the maintenance performance while properly processing the involved uncertainties has been proposed by the authors in [21], [40], where the uncertainty in the model parameters is represented by resorting to the DSTE theory. In particular, the main differences between this latter approach and that presented in this work are outlined in [21], with respect to the information elicited, the readability of the output, the computational effort required and the possibility of passing from the PT-FRV-based method to the DSTE-based one.…”

A novel risk-based analysis for the production system under epistemic uncertainty

“…The different type of available information calls for the development of a different method to represent and propagate the associated uncertainty. Resorting to probability theory to address this issue is questionable (Baraldi et al 2013). The use of evidence theory started with Dempster's work by description of the accounting principles of the upper and lower probabilities, and mathematical theory of evidence was defined precisely by Shafer (1976).…”